Vertebral Fracture Quantification

1. A method of deriving an estimate of the extent of fracture in a vertebra by processing an image of part of a spine, comprising the steps of: segmenting the image of at least two vertebrae in the spine to obtain data representative of the shape and size of each of the vertebrae; reconstructing an approximation of the shape of a first of the vertebrae by: comparing the shape and size data obtained for a second of the two vertebrae with a mathematical shape and size model of at least the same two vertebrae of an unfractured spine; adapting the second vertebra in the model to fit the second imaged vertebra with consequent changes to the remainder of the model; and predicting the unfractured shape and size of the first vertebra using the first vertebra of the adapted model; comparing the shape and size of the first vertebra as imaged with the predicted unfractured shape and size of the first vertebra; and computing the difference between the respective images to obtain a result representative of the extent of fracture in the first vertebra.

2. A method as claimed in claim 1 , further comprising the steps of: reconstructing an approximation of the shape and size of the second vertebra by: comparing the data obtained for the first vertebra with the mathematical model; adapting the first vertebra in the model to fit the first imaged vertebra with consequent changes to the remainder of the model; and predicting the unfractured shape and size of the second vertebra using the second vertebra of the adapted model; comparing the shape and size of the second vertebra as imaged with the predicted unfractured shape and size of the second vertebra; and computing the differences between the respective images of the first and second vertebrae to obtain a result representative of the extent of fracture in the set of vertebrae.

3. A method as claimed in claim 1 or claim 2 , further comprising the steps of: segmenting the images of a third to nth vertebrae in the image of the spine to obtain data representative of the shape and size of the third to nth vertebrae; reconstructing further approximations of the shape and size of the first vertebrae by: comparing the data for the third vertebra with a mathematical model of at least the same n vertebrae of an unfractured spine; adapting the third vertebra in the model to fit the third imaged vertebra with consequent changes to the remainder of the model; and predicting the unfractured shape and size of the first vertebra using the first vertebra of the adapted model; repeating the steps of adapting and predicting for each of the fourth to the nth vertebra to obtain n−1 predictions of the unfractured shape and size of the first vertebra; processing the n−1 predictions of the unfractured shape and size of the first vertebra to obtain a single prediction of the unfractured shape and size of the first vertebra representative of the n−1 predictions; comparing the shape and size of the first vertebra as imaged with the single predicted unfractured shape and size of the first vertebra; and computing the difference between the respective images to obtain a result representative of the extent of fracture in the set of vertebrae.

4. A method as claimed in claim 3 , further comprising the steps of: reconstructing an approximation of the shape and size of each of the n vertebrae by: comparing each of the n vertebrae in turn with the mathematical model; adapting the mathematical model to each of the other n−1 vertebrae in turn to obtain n−1 predictions of each of the vertebrae; processing the n−1 predictions of the unfractured shape and size of each vertebrae to obtain a single prediction of the unfractured shape and size of each of the n vertebrae; comparing the shape and size of each vertebra as imaged with the single predicted unfractured shape and size of the vertebra; and computing the difference between the respective images to obtain a result representative of the extent of fracture in the set of vertebrae.

5. A non-transitory computer-readable medium encoded with an instruction set for deriving an estimate of the extent of fracture in a vertebra of a processed image of part of a spine, the instruction set comprising instructions for: segmenting the image of at least two vertebrae in the spine to obtain data representative of the shape and size of each of the vertebrae; reconstructing an approximation of the shape of a first of the vertebrae including: comparing the shape and size data obtained for a second of the two vertebrae with a mathematical shape and size model of at least the same two vertebrae of an unfractured spine; adapting the second vertebra in the model to fit the second imaged vertebra with consequent changes to the remainder of the model; and predicting the unfractured shape and size of the first vertebra using the first vertebra of the adapted model; comparing the shape and size of the first vertebra as imaged with the predicted unfractured shape and size of the first vertebra; and computing the difference between the respective images to obtain a result representative of the extent of fracture in the first vertebra.

6. A non-transitory computer-readable medium encoded with an instruction set as claimed in claim 5 , further comprising instructions for: reconstructing an approximation of the shape and size of the second vertebra by: comparing the data obtained for the first vertebra with the mathematical model; adapting the first vertebra in the model to fit the first imaged vertebra with consequent changes to the remainder of the model; and predicting the unfractured shape and size of the second vertebra using the second vertebra of the adapted model; comparing the shape and size of the second vertebra as imaged with the unfractured predicted shape and size of the second vertebra; and computing the differences between the respective images of the first and second vertebrae to obtain a result representative of the extent of fracture in the set of vertebrae.

7. A non-transitory computer-readable medium encoded with an instruction set as claimed in claim 5 or claim 6 , further comprising instructions for: segmenting the images of a third to nth vertebrae in the image of the spine to obtain data representative of the shape and size of the third to nth vertebrae; reconstructing further approximations of the shape and size of the first vertebrae including: comparing the data for the third vertebra with a mathematical model of at least the same n vertebrae of an unfractured spine; adapting the third vertebra in the model to fit the third imaged vertebra with consequent changes to the remainder of the model; and predicting the unfractured shape and size of the first vertebra using the first vertebra of the adapted model; repeating the steps of adapting and predicting for each of the fourth to the nth vertebra to obtain n−1 predictions of the unfractured shape and size of the first vertebra; processing the n−1 predictions of the shape and size of the first vertebra to obtain a single prediction of the unfractured shape and size of the first vertebra representative of the n−1 predictions; comparing the shape and size of the first vertebra as imaged with the single predicted unfractured shape and size of the vertebra; and computing the difference between the respective images to obtain a result representative of the extent of fracture in the set of vertebrae.

8. A non-transitory computer-readable medium encoded with an instruction set as claimed in claim 7 , further comprising instructions for: reconstructing an approximation of the shape and size of each of the n vertebrae including: comparing each of the n vertebrae in turn with the mathematical model; adapting the mathematical model to each of the n−1 vertebrae in turn to obtain n−1 predictions of each of the vertebrae; processing the n−1 predictions of the shape and size of each vertebrae to obtain a single prediction of the unfractured shape and size of each of the n vertebrae; comparing the shape and size of each vertebra as imaged with the single predicted unfractured shape and size of the vertebra; and computing the difference between the respective images to obtain a result representative of the extent of fracture in the set of vertebrae.

9. A data processor for deriving an estimate of the extent of fracture in a vertebra of a processed image of part of a spine, having stored therein an instruction set comprising instructions for: segmenting the image of at least two vertebrae in the spine to obtain data representative of the shape and size of each of the vertebrae; reconstructing an approximation of the shape of a first of the vertebrae including: comparing the shape and size data obtained for a second of the two vertebrae with a mathematical shape and size model of at least the same two vertebrae of an unfractured spine; adapting the second vertebra in the model to fit the second imaged vertebra with consequent changes to the remainder of the model; and predicting the unfractured shape and size of the first vertebra using the first vertebra of the adapted model; comparing the shape and size of the first vertebra as imaged with the predicted unfractured shape and size of the first vertebra; and computing the difference between the respective images to obtain a result representative of the extent of fracture in the first vertebra.

10. A data processor as claimed in claim 9 , having stored therein further instructions for: reconstructing an approximation of the shape and size of the second vertebra by: comparing the data obtained for the first vertebra with the mathematical model; adapting the first vertebra in the model to fit the first imaged vertebra with consequent changes to the remainder of the model; and predicting the unfractured shape and size of the second vertebra using the second vertebra of the adapted model; comparing the shape and size of the second vertebra as imaged with the predicted unfractured shape and size of the second vertebra; and computing the differences between the respective images of the first and second vertebrae to obtain a result representative of the extent of fracture in the set of vertebrae.

11. A data processor as claimed in claim 9 or claim 10 , having stored therein further instructions for: segmenting the images of a third to nth vertebrae in the image of the spine to obtain data representative of the shape and size of the third to nth vertebrae; reconstructing further approximations of the shape and size of the first vertebrae including: comparing the data for the third vertebra with a mathematical model of at least the same n vertebrae of an unfractured spine; adapting the third vertebra in the model to fit the third imaged vertebra with consequent changes to the remainder of the model; and predicting the unfractured shape and size of the first vertebra using the first vertebra of the adapted model; repeating the steps of adapting and predicting for each of the fourth to the nth vertebra to obtain n−1 predictions of the unfractured shape and size of the first vertebra; processing the n−1 predictions of the shape and size of the first vertebra to obtain a single prediction of the shape and size of the first vertebra representative of the n−1 predictions; comparing the shape and size of the first vertebra as imaged with the single predicted unfractured shape and size of the vertebra; and computing the difference between the respective images to obtain a result representative of the extent of fracture in the set, of vertebrae.

12. A data processor as claimed in claim 11 , having stored therein further instructions for: reconstructing an approximation of the shape and size of each of the n vertebrae including: comparing each of the n vertebrae in turn with the mathematical model; adapting the mathematical model to each of the n−1 vertebrae in turn to obtain n−1 predictions of each of the vertebrae; processing the n−1 predictions of the shape and size of each vertebrae to obtain a single prediction of the unfractured shape and size of each of the n vertebrae; comparing the shape and size of each vertebra as imaged with the single predicted unfractured shape and size of the vertebra; and computing the difference between the respective images to obtain a result representative of the extent of fracture in the set of vertebrae.